Synthetic Process Mining Log Generator

Define a simple process model, then generate a CSV event log for quick process-mining experiments.

Global Settings

Timestamps for each case start near this value and increase per activity.

Resources

NameRemove

Resources are assigned to activities. If an activity uses "Random", one of these will be picked.

Working Time (Calendar Constraints)

From
To

Event timestamps will be clamped into these working windows.

Activities

Activity name Avg duration (mins) Offset min (mins) Offset max (mins) Resource assignment Rework settings Remove

Duration = avg minutes + random offset (min to max). Rework can loop back to the chosen target before following outgoing transitions.

Conditional Delays (Optional)

Activity Field Operator Value Delay amount Delay unit Remove

Add extra wait before an activity when the condition matches (case or event attributes). Multiple matched rules are summed.

Flow Definition

From activity Next activity Probability Remove

Probabilities leaving an activity should sum to 1. Use the special "END" target to finish a case.

Case System (case-level source)

System name Probability Remove

Pick which system the case runs in, using the configured probabilities.

Case Attributes

Attribute name Type Definition Remove

Adds metadata or dimensions per case. Categorical values are picked randomly; numeric values use a random value within the range.

Use the floating toolbar to generate and download logs.

Created by Matthias Gielkens with Codex

About

Organizations often want to explore process mining, run a proof of concept, or compare different process intelligence tools. But there is a recurring blocker: getting suitable event log data.

Real operational data is usually difficult to extract, contains sensitive information, and requires significant preparation before it can be used. This slows down experimentation, onboarding, training, and early prototyping.

The Synthetic Process Mining Log Generator solves this by letting business users sketch a lightweight process model and instantly generate realistic CSV event logs that mimic real-world behaviors — without accessing any confidential systems.

What this project enables

  • Fast experimentation: Build and adjust simple process flows and immediately produce logs for discovery, conformance checking, and performance analysis.
  • Training & demos: Generate clean sample datasets to teach process mining concepts or demonstrate tooling.
  • Early project shaping: Model a “to-be” or illustrative “as-is” process when real data isn’t yet available.
  • Vendor evaluation: Test multiple process mining platforms with identical synthetic logs.

Why this is valuable for business users

  • No dependence on IT data extraction
  • Safe for experimentation — no personal or sensitive data
  • Quick to iterate — change your process design and regenerate in seconds
  • Ideal for proofs of concept where clean data is needed quickly
  • Supports storytelling — demonstrate the value of process mining using relatable, customizable scenarios

How to use it

  1. Define the process structure. Outline activities, decision branches, sequences, and end points. Optionally describe resources/roles, duration expectations, and execution type (manual, system, robotic).
  2. Add realistic behavior. Enrich with rework or loops (with probability and limits), conditional delays (e.g., specific customer segments taking longer), and case attributes such as region, customer type, risk level, or product.
  3. Configure volumes. Set how many cases you want and how many events each should generate. The system validates probabilities, flow completeness, and duration logic automatically.
  4. Generate & preview. Simulate the process to create timestamped events, preview the first rows, and download CSV event logs and case-level tables compatible with standard process mining tools.
  5. Import and analyze. Load the synthetic logs into your preferred platform (e.g., Celonis, UIPath Process Mining, Apromore, PAFnow, Disco) to explore variants, bottlenecks, lead times, rework, and compliance patterns.